s c a l e When Hardware Attacks Hardwear.io 27 September 2019 1 - - PowerPoint PPT Presentation

1

When Hardware Attacks

Hardwear.io 27 September 2019

scale

2

Attack exploitation space: time vs distance

protocol software

Local Fast Slow Remote

key brute force side channel fault injection physical relay attack

Hardware attacks require:

• hardware vulnerabilities, or
• hardware changes to target

mitm Hardware attacks

3

𝑞 = 𝑜 ∗ 𝑤 − 𝑑𝑤 − 𝑑𝑔

p = profit v = value n = replications cv = variable costs cf = fixed costs

Attacker business case

3

4

1.EMV Man-in-the-Middle Hardware attack to bypass PIN verification of stolen payment cards 2.Retail hack Network penetration attack to retrieve cardholder credentials 3.Card sharing Relay attack to avoid paying TV subscription fees

Let’s analyze some known attacks

4

5

EMV Man-in-the-Middle (1)

5

Source: https://www.cl.cam.ac.uk/research/security/banking/nopin/

6

EMV Man-in-the-Middle (2)

6

7

Retail hack

7

8

• Pay-TV decoders use smart

cards to control video access

• Subscription is in smart card

Card sharing (1)

8

9

• Pay-TV decoders use smart cards to

control video access

• Subscription is in smart card
• Distribution of session keys avoids

need for individual subscriptions

Card sharing (2)

9

10

Attack Fixed Cost Variable Cost Value Replications Profit EMV MitM € 30K € 100 € 500 100 € 10 K Retail hack € 20K € 1 € 25 10K € 220 K Card sharing € 10K € 10 € 100 1M € 90 M

Example attack business cases

10 Replications are key, but how is that bounded?

• Application size (e.g. #potential victims)
• Detection & mitigation
• Replication effort

To determine scalability, we need to quantify the replication effort

11

Attack phases and cost

Identification Exploitation What it is finding a vulnerability replicate on target Frequency

• nce

repeated Speed How fast can we do this? Skill Required knowledge / experience Equipment Type of equipment Location Where is the attacker?

What parameters determine the attack cost? Fixed cost Variable cost

12

Attack parameters

Identification Exploitation

Vulnerability Hardware Software Hardware Software Speed slow slow slow fast Skill expert expert proficient layman Equipment specialized standard specialized none Location local near local remote

Scalable attack

Scalable attacks need software exploitation! What are typical attack parameters?

13 Defenders method Attackers method

How to find software vulnerabilities?

13

White-Box Black-Box Source Code Review Binary Analysis Fuzzing Model Based Testing Effectiveness Most vulnerabilities are found white-box style!

14

Finding vulnerabilities in source code

Software packages typically

• vary between 10 and 10,000 KLoC
• have 0.1 up to 10 vulnerabilities per KLoC

 All products have software vulnerabilities Manual source code review performs at 100 LoC/hr  Finding a vulnerability in source code may take just one day

14

15

Binary analysis

16

Disassemble

17

Flow analysis

18

• Given the widespread presence of vulnerabilities there is an

increasing desire to mitigate risk

• Finding software vulnerabilities gets more difficult without

access to source/binary code

• Access to device software is increasingly restricted:
• PC software used to be accessible (e.g. exe files)
• Smart phone software is only visible for root
• Set-Top-Box software is hidden, and encrypted in transit
• How to attack a product protected with software encryption?

Software vulnerability hiding

internal

18

19

Attacking encrypted software

Hardware attack offers two-step alternative:

• 1. Break software confidentiality
• 2. White-box binary analysis exposes logical vulnerability

19

Binary analysis exposes logical vulnerability Exploitation yields runtime control Start Black-Box penetration testing exposes logical vulnerability Start Exploitation yields runtime control Hardware attack breaks software confidentiality Start Binary analysis exposes logical vulnerability Exploitation yields runtime control

Encrypted software hides binary code Black-Box penetration testing very inefficient

20

Source: http://www.fredericb.info/2016/10/amlogic-s905-soc-bypassing-not-so.html

Design flaw in Pay-TV SoC

21

Secure boot chain broken by backdoor

Attacker used hardware weakness to dump Boot Loader image

Restricted

21

22

Boot Loader header analysis

struct aml_img_header { // 64 bytes unsigned char magic[4];// "@AML" uint32_t total_len; uint8_t header_len; uint8_t unk_x9; uint8_t unk_xA; uint8_t unk_xB; uint32_t unk_xC; uint32_t sig_type; uint32_t sig_offset; uint32_t sig_size; uint32_t data_offset; uint32_t unk_x20; uint32_t cert_offset; uint32_t cert_size; uint32_t data_len; uint32_t unk_x30; uint32_t code_offset; uint32_t code_len; uint32_t unk_x3C; } aml_img_header_t;

22 sig_type provides backdoor that bypasses verification

Select Certificate Go Get key Public key Code Hash Signature Verified Sig Verify Hashed code Compare Stop

23

Conclusions

• Scalable attacks need software exploitation

– Hardware attacks are laborious – Software vulnerabilities are ubiquitous – Software exploits are easy to replicate

• Software encryption is inevitable for security

– Binary analysis very successful in identifying vulnerabilities – Increasing number of products use encrypted software

• Hardware attacks are scalable when

– Software is encrypted – Shallow bugs (detectable black-box style) are absent – Used in the identification step to extract software – Deep software vulnerabilities are present

23

driving your security forward

Riscure B.V.

Frontier Building, Delftechpark 49 2628 XJ Delft The Netherlands Phone: +31 15 251 40 90 www.riscure.com

Riscure North America

550 Kearny St., Suite 330 San Francisco, CA 94108 USA Phone: +1 650 646 99 79 inforequest@riscure.com

Riscure China

Room 2030-31, No. 989, ChangleRoad, Shanghai 200031 China Phone: +86 21 5117 5435 inforcn@riscure.com

Riscure is hiring, visit https://www.riscure.com/careers/

Questions? contact: Marc Witteman witteman@riscure.com